Mailman 3 January 2011 - NumPy-Discussion

Getting a clone copy of the NumPy repository.
by Hector troy 16 Jan '11

16 Jan '11

Hello everyone, I am a newbie on this open source world, and sincerely trying to make contribution to the development of Numpy. I was trying to learn about making patches from http://docs.scipy.org/doc/numpy/dev/gitwash/patching.html but unable to get the clone of Numpy repository. In the terminal error massage shown is - $hector@hector:~$ sudo git clone git://github.com/numpy/numpy.git [sudo] password for hector: Initialized empty Git repository in /home/hector/numpy/.git/ github.com[0: 207.97.227.239]: errno=Connection timed out fatal: unable to connect a socket (Connection timed out) $hector@hector:~$ My internet is quite good at the moment but unable to understand why am I getting this error. Any help regarding this will be extremely helpful and encouraging for me. Thanking you in anticipation. -- -Regards Hector Whenever you think you can or you can't, in either way you are right.

2 2

float conversion
by Alex Ter-Sarkissov 16 Jan '11

16 Jan '11

hi every1, I got the following issue: I wrote a function that converts binary strings into a decimal value (binary expansion). When I write type(x) to find out the type of the value I get NoneType. Therefore I can't convert it into anything else, such as float numbers, since command float(x) returns TypeError: float() argument must be a string or a number Any ideas what to do with this (e.g. convert to floating numbers)? cheers, Alex

2 1

Is Savitzky-Golay filter provided by cookbook different from matlab's?
by Hexyl Chan 15 Jan '11

15 Jan '11

I have used the Savitzky-Golay filter provided by cookbook, but the result is different from the one given by matlab function sgolayfilt(x,2,15). I attached the numpy code below: #! /usr/bin/env python # -*- coding:utf-8 -*- import numpy as np def savitzky_golay(y, window_size, order, deriv=0): r"""Smooth (and optionally differentiate) data with a Savitzky-Golay filter. The Savitzky-Golay filter removes high frequency noise from data. It has the advantage of preserving the original shape and features of the signal better than other types of filtering approaches, such as moving averages techhniques. Parameters ---------- y : array_like, shape (N,) the values of the time history of the signal. window_size : int the length of the window. Must be an odd integer number. order : int the order of the polynomial used in the filtering. Must be less then `window_size` - 1. deriv: int the order of the derivative to compute (default = 0 means only smoothing) Returns ------- ys : ndarray, shape (N) the smoothed signal (or it's n-th derivative). Notes ----- The Savitzky-Golay is a type of low-pass filter, particularly suited for smoothing noisy data. The main idea behind this approach is to make for each point a least-square fit with a polynomial of high order over a odd-sized window centered at the point. Examples -------- t = np.linspace(-4, 4, 500) y = np.exp( -t**2 ) + np.random.normal(0, 0.05, t.shape) ysg = savitzky_golay(y, window_size=31, order=4) import matplotlib.pyplot as plt plt.plot(t, y, label='Noisy signal') plt.plot(t, np.exp(-t**2), 'k', lw=1.5, label='Original signal') plt.plot(t, ysg, 'r', label='Filtered signal') plt.legend() plt.show() References ---------- .. [1] A. Savitzky, M. J. E. Golay, Smoothing and Differentiation of Data by Simplified Least Squares Procedures. Analytical Chemistry, 1964, 36 (8), pp 1627-1639. .. [2] Numerical Recipes 3rd Edition: The Art of Scientific Computing W.H. Press, S.A. Teukolsky, W.T. Vetterling, B.P. Flannery Cambridge University Press ISBN-13: 9780521880688 """ try: window_size = np.abs(np.int(window_size)) order = np.abs(np.int(order)) except ValueError, msg: raise ValueError("window_size and order have to be of type int") if window_size % 2 != 1 or window_size < 1: raise TypeError("window_size size must be a positive odd number") if window_size < order + 2: raise TypeError("window_size is too small for the polynomials order") order_range = range(order+1) half_window = (window_size -1) // 2 # precompute coefficients b = np.mat([[k**i for i in order_range] for k in range(-half_window, half_window+1)]) m = np.linalg.pinv(b).A[deriv] # pad the signal at the extremes with # values taken from the signal itself firstvals = y[0] - np.abs( y[1:half_window+1][::-1] - y[0] ) lastvals = y[-1] + np.abs(y[-half_window-1:-1][::-1] - y[-1]) y = np.concatenate((firstvals, y, lastvals)) return np.convolve( m, y, mode='valid') if __name__ == "__main__": rrstype = np.dtype({'names':['wavelenth', 'rrs'], 'formats':['f','f']}) record_list = [] record_file = open("200607NO1.txt") for line in record_file: item_list = line.strip().split('\t') record_list.append((item_list[0], item_list[1])) record_file.close() record_array = np.array(record_list,dtype=rrstype) rrs_array = record_array["rrs"] rrs_array_2 = savitzky_golay(rrs_array, 15, 2) rrs_array_2.tofile("savitzky-golay-result.txt", "\n")

1 0

special function xlogy with 0log0=0
by josef.pktd＠gmail.com 15 Jan '11

15 Jan '11

I'm trying to fix again a case with x*log(y) which was a converted log(0**0), where x and y should broadcast. Is it possible to get a special function for this. It shows up very often, and any cheap fix, e.g. term = x*np.log(y) term[(x==0)*(y==0)] = 0 raises a warning (which I now also see) x*log(y+1e-300) would be easier, but I don't know whether there are no cases where numerical precision deteriorates. In another version, Skipper clipped y What's the best way? Josef

1 0

isposinf returns array, isinf doesn't
by josef.pktd＠gmail.com 14 Jan '11

14 Jan '11

maybe just cosmetic, I just found this >>> stats.poisson.b 1.#INF >>> np.isinf(stats.poisson.b) True >>> np.isinf(-stats.poisson.b) True >>> np.isposinf(stats.poisson.b) array(True, dtype=bool) >>> np.isneginf(stats.poisson.b) array(False, dtype=bool) >>> np.isneginf(-stats.poisson.b) array(True, dtype=bool) but shape is the same >>> np.isneginf(stats.poisson.b).shape () >>> np.isinf(stats.poisson.b).shape () >>> type(np.isneginf(stats.poisson.b)) <type 'numpy.ndarray'> >>> type(np.isinf(stats.poisson.b)) <type 'numpy.bool_'> Josef

1 0

NaN value processing in weave.inline code
by Joon Ro 14 Jan '11

14 Jan '11

Hi, I was wondering if it is possible to process (in if statement - check if the given value is NaN) numpy NaN value inside the weave.inline c code. testcode = ''' if (test(0)) { return_val = test(0); } ''' err = weave.inline(testcode, ['test'], type_converters = converters.blitz, force = 0, verbose = 1) with test(0) = nan returns err = nan correctly, but I don't know how to check the nan value inside the c inline c code. Is there any way I can get similar functionality as isnan? Thank you, Joon --

2 2

Is python 3 supported or not?
by David Cortesi 14 Jan '11

14 Jan '11

I asked about getting numpy to install on OS X with Activestate Python 3. I got thoughtful & responsive replies from three of you, many thanks to all! I am sad that the consistent message was, "forget it." Chris said, "...look for a binary for 3.* -- I'm not sure it exists, though." Sebastian said, "the simple answer you might be looking for is: it's easier to stay with Python 2.x..." Ralf said, "…unless you have a specific need/desire to use 3.1/3.2 I'd suggest staying with 2.6 or 2.7 from python.org for now..." I would like to point out that the wikipedia article on numpy says, "The release version 1.5 of NumPy is compatible with Python versions 2.4–2.7 and Python 3," citing the release note of september 2010, which itself opens with the following lines: > Highlights > > Python 3 compatibility > > This is the first NumPy release which is compatible with Python 3. There is an obvious disconnect here. Is it or isn't it? This is an important question because of the large number of packages at PyPI that depend on numpy. Numpy is a major gateway, or bottleneck, on the way to Python 3. I came looking for numpy because I want to work with an audio package, and all the audio packages at PyPI seem to have numpy dependencies. Ditto the packages for dealing with FITS data format, etc. etc. As to using Activestate's versus python.org's distro, *regardless* of which I use, the package will end up located in /Library/Frameworks/Python.framework/*. It will not be installed in /System/Library/etc. as the Apple distribution is; but it WILL be located at a known location with an executable named Python under Versions/Current. Not that it matters, but the reason I'm using Activestate is because I also needed their up to date version of Tcl/Tk, and python.org python3 wouldn't work with that. As to why I'm using Python 3, it's because I'm starting a new project with no prior dependencies and want the current and future language -- which is now TWO FRAKKIN' YEARS OLD! -- but that's a rant for another time. Thanks again for your attention, Dave Cortesi

6 6

Any idea to run the dot-product on many arrays
by EMMEL Thomas 13 Jan '11

13 Jan '11

Hi, I need to rotate many vectors (x,y,z) with a given rotation matrix (3x3). I can always do for v in vectors: tv += np.dot(mat, v) where mat is my fixed matrix (or array of arrays) and v is a single array. Is there any efficient way to use an array of vectors to do the transfomation for all of these vectors at once? Kind regards Thomas This email and any attachments are intended solely for the use of the individual or entity to whom it is addressed and may be confidential and/or privileged. If you are not one of the named recipients or have received this email in error, (i) you should not read, disclose, or copy it, (ii) please notify sender of your receipt by reply email and delete this email and all attachments, (iii) Dassault Systemes does not accept or assume any liability or responsibility for any use of or reliance on this email.For other languages, go to http://www.3ds.com/terms/email-disclaimer.

2 1

Numpy 1.5.1 - Mac - with Activestate Python 3
by David Cortesi 13 Jan '11

13 Jan '11

I have installed ActiveState's Python 3 packages on Mac OS X 10.6.6. There exists: /Library/Frameworks/Python.framework/Versions/Current/Python* When I run the Mac OS installer it shows all disks as ineligible and the error message, "numpy 1.5.1 can't be installed on this disk. numpy requires System Python 2.6 to install." What can I do to persuade numpy to install? Must I build it from source to get it to use Python 3? Sorry for the Noob question, Dave Cortesi

4 3

Output dtype
by Keith Goodman 12 Jan '11

12 Jan '11

>From the np.median doc string: "If the input contains integers, or floats of smaller precision than 64, then the output data-type is float64." >> arr = np.array([[0,1,2,3,4,5]], dtype='float32') >> np.median(arr, axis=0).dtype dtype('float32') >> np.median(arr, axis=1).dtype dtype('float32') >> np.median(arr, axis=None).dtype dtype('float64') So the output doesn't agree with the doc string. What is the desired dtype of the accumulator and the output for when the input dtype is less than float64? Should it depend on axis? I'm trying to duplicate the behavior of np.median (and other numpy/scipy functions) in the Bottleneck package and am running into a few corner cases while unit testing. Here's another one: >> np.sum([np.nan]).dtype dtype('float64') >> np.nansum([1,np.nan]).dtype dtype('float64') >> np.nansum([np.nan]).dtype <snip> AttributeError: 'float' object has no attribute 'dtype' I just duplicated the numpy behavior for that one since it was easy to do.

2 6